REVIEW Open Access

An evaluation of HHV-6 as an etiologicagent in Hodgkin lymphoma and braincancer using IARC criteria for oncogenicityMichael J. Wells1 , Steven Jacobson2 and Paul H. Levine3*

Abstract

Background: Human herpesvirus-6 (HHV-6) is a ubiquitous double-stranded DNA virus that can cause roseolainfantum, encephalitis, and seizure disorders. Several studies have shown an association between HHV-6 and cancerbut confirmation of an etiologic role is lacking. We reviewed the criteria for viral causation of cancer used by TheInternational Agency for Research on Cancer (IARC) for six oncogenic viruses and applied criteria to publishedreports of HHV-6 and its association with Hodgkin lymphoma and brain tumors.

Methods: Our major criteria for oncogenicity were finding evidence of the virus in every tumor cell and preventionof the tumor by an antiviral vaccine. Our six minor criteria included: 1) suggestive serologic correlation, such ashigher virus antibody levels in cases compared to controls; 2) evidence of the virus in some but not all tumor cells,and 3) time space clustering. We focused on Epstein-Barr virus (EBV) as the primary virus for comparison as HHV-6and EBV are both Herpesviridae, ubiquitous infections, and EBV is well-accepted as a human oncovirus. Particularattention was given to Hodgkin lymphoma (HL) and brain cancer as these malignancies have been the moststudied.

Results: No studies reported HHV-6 satisfying either of the major criteria for oncogenicity. Of the minor criteriaused by IARC, serologic studies have been paramount in supporting EBV as an oncogenic agent in all EBV-associated tumors, but not for HHV-6 in HL or brain cancer. Clustering of cases was suggestive for both HL andbrain cancer and medical intervention suggested by longer survival in patients treated with antiviral agents wasreported for brain cancer.

Conclusion: There is insufficient evidence to indicate HHV-6 is an etiologic agent with respect to HL and braincancers. We suggest that methods demonstrating EBV oncogenicity be applied to HHV-6. It is important that onestudy has found HHV-6 in all cancer cells in oral cancer in a region with elevated HHV-6 antibodies and thereforeHHV-6 can still be considered a possible human oncogenic virus.

Keywords: HHV-6, Human oncogenic virus, Hodgkin’s lymphoma, Hodgkin lymphoma, Hodgkin’s disease, Braincancer, Oral cancer

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

* Correspondence: paulhlevine@earthlink.net3College of Public Health, University of Nebraska, 984355 Medical Center,Omaha, NE 68198, USAFull list of author information is available at the end of the article

Wells et al. Infectious Agents and Cancer (2019) 14:31 https://doi.org/10.1186/s13027-019-0248-3

BackgroundHuman herpesvirus-6 (HHV-6) is a ubiquitous infectiousagent that is the primary cause of roseola infantum, anacute childhood illness characteristically presenting witha high fever followed by a generalized rash. Other mani-festations of primary infection with HHV-6 have beeninvestigated [1] and current research is focusing on itsneurotropic properties suggesting a link to encephalitis,seizure disorders, Alzheimer’s disease [2], and perhapsmultiple sclerosis [3, 4]. The association of HHV-6 withcancer dates back to the first report of its isolation frompatients with hematologic malignancies. Initially de-scribed as a human B-cell lymphotropic virus (HBLV)[5], HHV-6 has now been shown to be more pleotropic,growing well in CD4+ T lymphocytes [6–8] and alsoable to replicate in macrophages and fibroblasts amongother cells [1].Initially, two strains of HHV-6, designated as HHV-6A

and HHV-6B, were defined with differences noted ingeographic distribution, primary reservoir areas, and as-sociated diseases in humans [1, 3, 9]. More recently,these two agents have been classified as distinct viruses[9]. Both are closely related Herpesviridiae in the betasubfamily of Herpesviridiae, along with cytomegalovirusand HHV-7. In contrast, Epstein-Barr virus (EBV), an-other member of the Herpesviridiae family, is classifiedin the gamma herpesvirus subfamily. The two HHV-6viruses have a unique form of latency. Unlike EBV, theydo not form episomes but rather establish latency by in-tegrating near the telomere of the chromosome [10].HHV-6B appears to be spread primarily through saliva

[3], although it has been detected in stool samples [11, 12]and vaginal secretions [13]. It is commonly transmittedfrom mother-to-infant. HHV-6A is more prevalent inadults compared to children, and to date has not been de-finitively associated with human disease, unlike HHV-6B[1]. The pattern of spread of different infectious oncogenicagents has been important in indicating the relationship tohuman cancer. Human T-cell lymphotropic Virus Type-I(HTLV-I), for example, is highly cell-associated and notreadily transmissible. There is a high prevalence of thisvirus in only a few areas, particularly Japan and theCaribbean [14–16]. Therefore, the strong geographiccorrelation between the diseases resulting from infec-tion with this virus, such as adult T-cell leukemia/lymphoma (ATLL) and HTLV-I associated myelopathy(HAM), constitutes strong support for the etiologicalrole of HTLV-I in those diseases.As molecular techniques have advanced, the criteria

for determining whether an infectious agent causes cancerhave changed. The classic criteria of disease causation islong-standing and includes suggestions from Henleand Koch [17–19], Bradford Hill [20], Rivers [21], andFredericks and Relman [22], who focused on detection of

the virus by in situ methods in each of the tumor cells.Under this direct hit model, the agent transforms an ini-tially healthy cell into a malignant cell and thereafter per-sists in all of the subsequent tumor cells. Moore andChang [23] recently used newly developed molecular tech-niques to implicate HHV-8, a gamma herpesvirus, as thecause of Kaposi’s sarcoma, and the Merkel Cell tumorvirus (MCV), a polyomavirus, as the cause of Merkel cellcarcinomas. However, with all of these agents, it is unclearif the continued presence of the virus is required to main-tain the tumor once oncogenesis is initiated.Other mechanisms of oncogenesis have been described,

primarily through chronic inflammation which causes cel-lular proliferation. Hepatitis C virus (HCV) causes hepato-cellular carcinoma through the intermediary of hepaticcirrhosis; non-viral infectious agents also cause cancer viachronic inflammation (e.g. Helicobacter pylori and gastriccancer, Schistosoma haematobium and bladder cancer,and Opisthorchis viverrini and bile duct cancer) [24].As we have learned more about human oncogenic

agents, it is clear that some of the early criteria for dis-ease causation, such as specificity (a one-to-one relation-ship or singular causal agent causing a singular diseaseposed by Bradford Hill), do not apply to oncogenicviruses. Gastric cancer appears to have at least two in-fectious etiologic triggers, H. pylori [25, 26] and Epstein-Barr virus [27–29]. Hepatocellular cancer can be causedby HCV or Hepatitis B virus (HBV) through entirelydifferent mechanisms and can also occur with anychronic hepato-destructive process. Epstein-Barr virusis etiologically linked to several tumors: Hodgkinlymphoma (HL) in approximately 36% of cases [30],Burkitt lymphoma in the vast majority of cases in sub-Saharan Africa, non-keratinizing nasopharyngeal car-cinoma (NPC) in almost every case worldwide, as wellas other malignancies [27–29, 31, 32].Currently, the International Agency for Research on

Cancer (IARC) develops consensus panels and rates theoncogenic potential in three groups: Group 1—carcino-genic for humans; Group 2—sufficient data to concludeeither probably carcinogenic for humans (2A) or possiblycarcinogenic for humans (2B); and Group 3—not classifi-able as carcinogenic for humans, which can mean eitherno evidence favoring a link or no studies examining theissue of carcinogenicity. We reviewed the IARC criteriafor the six viruses that have been evaluated in IARCmonographs for their criteria for oncogenicity and appliedthese criteria to published reports of HHV-6 and its asso-ciation with Hodgkin lymphoma and brain tumors.

MethodsFirst, we listed the most important criteria for oncogen-icity identified by IARC for the Group 1 viral carcinogens.The six viruses classified as Group 1 carcinogens were

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 2 of 12

reviewed: Epstein-Barr virus (EBV) [33], Human Herpes-virus 8 or Kaposi’s sarcoma herpesvirus (HHV-8, KSHV)[34], Human T-lymphotropic virus Type I (HTLV-I) [35],Hepatitis B and C viruses (HBV and HCV) [33, 36, 37],and human papilloma virus (HPV) [33, 38, 39]. We didnot include MCV as it has not been classified in Group 1at this time, the latest review stating “There is limitedevidence in humans for carcinogenicity” [40]. Also, we didnot include the human immunodeficiency virus (HIV)because it is not directly oncogenic but rather causescancer through the associated immunodeficiency.We have chosen to divide the criteria for etiology into

two groups: major criteria (those that generally by them-selves are accepted as proving etiology and are invariablyaccompanied by a number of minor criteria) and minor cri-teria, (those that are frequently found in oncogenic virusescertified as Group 1 carcinogens (see below) by the IARC).For major criteria, we have chosen two findings that aregenerally accepted by viral oncologists as providing thestrongest evidence for oncogenicity: viral evidence in alltumor cells and tumor prevention by an antiviral vaccine.

Major criteriaVirus detection studiesStrong evidence of causality should begin with viral evi-dence in all tumor cells, which we designate as “diffuse”involvement to distinguish this from viral evidence insome tumor cells, which we designate as “restricted” in-volvement. Clearly, many cells can be infected by viruseswithout consequence, but this model assumes that ma-lignant transformation initially occurs in one cell, andthen this transformed cell carries viral information to itsprogeny. Because of this carriage, presence of the virusis constantly maintained in tumor cells. Examples of thiscriterion include EBV and undifferentiated nasopharyngealcarcinoma [41], EBV-associated gastric cancer [42–44], andEBV-associated Hodgkin lymphoma [45–48]. As a necessityfor this to be accepted as proof of causation, we added thecriterion of reproducibility with confirmation of persistentcarriage established by independent laboratories.

Tumor preventionAnother major criterion that is generally accepted andnoted by Evans and Mueller [49] is prevention of thetumor by blocking infection with the suspected etiologicagent. Thus far, only the HBV vaccine has met this cri-terion [50] although the HPV vaccine has effectively re-duced the number of pre-cancerous cervical lesions andwill likely reduce the incidence of cervical cancer in theimmunized populations [51].

Minor criteriaMinor criteria are important clues to oncogenicity andare often found to exist together in the same candidate

agent. The criteria include causative evidence in sero-logic and viral detection studies, biological gradient,clustering, medical interventions, and animal models.

Serologic studiesSerologic studies usually provide the first evidence foran association of an oncogenic virus and a disease.Numerous examples exist which have shown higherantibody titers in the associated disease than controls forall of the accepted oncogenic viruses. However, suchstudies can provide false leads, such as the initial focuson herpes simplex virus 2 (HSV-2) as the cause of cer-vical cancer [52]; both HSV-2 and the oncogenic strainsof human papillomavirus are sexually transmitted.

Case-control studiesCase-control studies provide some refinement since thecases are specifically matched by age, sex, and other rele-vant criteria, but as with HSV-2, if the confoundingagent is transmitted in the same way as the etiologicagent (as in this case HPV) careful matching is of littlehelp. Examples of useful case-control studies cited byIARC include reports on African Burkitt lymphoma (BL)patients by Henle et al. [53] and Klein et al. [54]. How-ever, higher antibody level might also occur when a pas-senger virus infecting the tumor replicates with thetumor even if it infects the tumor after it has started.Furthermore, tumor-induced cachexia might affect im-mune responses to a virus.

Cohort studiesCohort studies provide stronger evidence than case-control studies, being able to determine if subjects whodeveloped disease had higher antibody titers in the pre-disease sera. For EBV, a large prospective study [55]funded by the National Cancer Institute showed thathigh EBV antibody predicted the development of BL.Second, the BL patients whose EBV antibody titers werelower than the general population and whose tumorscould be studied had non-EBV associated BL, i.e., BLwithout the virus being present in the tumor cells. Simi-larly, Beasley et al. showed that Taiwanese subjects in-fected with HBV had a 223-fold higher risk of developinghepatocellular carcinoma than uninfected controls [56].This prospective study resulted in a HBV vaccine trialwhich eventually demonstrated that preventing infectionled to a decrease in mortality from the associated cancer[50], thus meeting one of our major criteria for causality.

Retrospective-prospective cohort studiesRetrospective-prospective cohort studies utilize alreadycollected serum samples and select sera to test based onthe outcome. For example, based on Shibata’s descrip-tion of EBV in a series of gastric carcinomas [42], we

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 3 of 12

selected sera from the Honolulu Heart Program cohortstudy of 8006 American men of Japanese ancestry andtested sera from 54 cases of non-lymphoepithelial gastriccarcinoma collected between 2.7 and 21.3 years (median14 years) prior to the diagnosis of cancer. This study,which essentially allowed the same approach as the pro-spectively collected sera from Ugandan children [55],showed higher antibody titers to EBV in the pre-EBVpositive gastric cancer patients compared to controls,whereas mean Helicobacter pylori antibody reactivitywas higher in the sera prior to development of EBV-negative gastric cancer patients and controls [57]. Coin-cidentally, antibody to HHV-6 was also tested in thisstudy and no difference between groups was noted.

Virus detection studies As noted above, viral evidencein every cell (diffuse involvement) and none in non-malignant cells is convincing evidence of viral etiology,and is considered a major criterion. However, virus de-tection in only some of the tumor cells (restricted in-volvement) is common. Tumors may contain infiltratingcells which can introduce the viruses they carry. Or theymay be heterogeneous (e.g., Hodgkin lymphoma is amixture of tumor and reactive cells). Thus, detectiondoes not necessarily signal etiology and all cells cannotbe expected to carry the virus, even if that virus is etio-logic. Moreover, some antibody reagents only detect thevirus in certain phases of replication in the malignantcell (e.g., early EBV studies and cancer where EBV viralcapsid, membrane and other antigens were found in-consistently). Uniform detection in all tumor cells re-quires the identification of EBV-encoded RNAs or(EBERs), which are expressed in all forms of EBV latency[45–48]. Thus, virus detection with restricted involvementis considered a minor criterion.

Biological gradient Both serologic and virus detectionstudies can provide important information as to biologicgradients. Overall, these studies analyze the relationshipbetween viral load and susceptibility to disease and/or severity of disease. There are numerous examplesof viral load paralleling disease burden, which can bereflected in antibody levels. Correlation of antibodytiters or viral load with the presence and particularlythe stage of disease can be helpful, as in BL andNPC where antibody titers to EBV-associated earlyantigen clearly parallel disease burden [58, 59]. Similarly,virus burden may signal higher tumor risk. With HTLV-I,virus load or elevated viral antibodies in the motherindicate the likelihood of the virus being passed onto the infant at birth or in the neonatal period,which is the primary risk factor for the developmentof ATLL [60].

Clustering These studies are particularly useful wherethe virus is not readily transmitted, such as with HTLV-Iwhich is highly cell-associated and is transmitted bybreast feeding, transfusion of blood cells or sexually. Thefinding that the HTLV-I endemic areas were preciselythe same as the ATLL endemic areas [14, 61] was thefirst convincing epidemiologic link of the virus with themalignancy. A similar finding is seen with classic Kapo-si’s sarcoma, which was shown to concentrate in south-ern Italy which has a higher prevalence of HHV-8antibodies than northern Italy [62]. This criterion hasnot been used for HHV-6 since it is a ubiquitous virusworldwide but geographic differences in titers in differ-ent populations have been noted [63].

Medical interventions If it is demonstrated that anti-viral treatment results in decreasing the virus burdenresulting in reduced pathogenesis and subsequent can-cer, this could be a clue that the virus load is importantin cancer etiology. Currently, this approach is promisingfor antiviral treatment of HCV and HBV to reduce theincidence of hepatocellular carcinoma. Reduction of theburdens of bacterial H. pylori and parasites such as O.viverrini are being employed to reduce the incidence ofgastric cancer and bile duct cancer respectively in en-demic areas.

Animal models In some instances, oncogenicity is dem-onstrated by inoculating animals with the human virusand producing a similar tumor, as suggested by the pro-duction of leukemic tumors in immune-defective mice in-oculated with HTLV-1-containing cells [64]. More often,a naturally occurring animal tumor is found to be pro-duced by a virus that resembles the human tumor. Earlystudies investigating the oncogenicity of EBV for humansrelied on observations of Marek’s disease, a lymphoma ofchickens produced by the Marek’ disease herpes viruswhich closely resembles EBV [65]. Also, naturally occur-ring feline leukemia virus produces diseases in cats thathave been considered a useful model for HTLV-I-associated ATLL in humans [66].With these major and minor criteria in mind, we eval-

uated the above criteria in the two human malignanciesthat have had the most data suggesting an etiologic rolefor HHV-6, the nodular sclerosis form of Hodgkinlymphoma and brain cancer. We attempted to identifyall peer-reviewed manuscripts published on the associ-ation of HHV-6 with Hodgkin lymphoma or brain can-cer. A systematic search was performed on PubMedusing combinations of the search terms “human herpes-virus 6, human herpesvirus-6, HHV6, HHV-6” and“Hodgkin’s lymphoma, Hodgkin lymphoma, Hodgkin’sdisease, Hodgkin disease,” or “brain cancer, brain tumor,brain tumors, glioma,” for a total of 36 separate searches. In

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 4 of 12

addition, several review articles on HHV-6 and Hodgkinlymphoma or brain cancer [67–69] were reviewed for iden-tification of relevant manuscripts that may not have beenidentified in our search.

ResultsThus far, no human tumor has been linked to HHV-6by either of our major criteria. The status of the minorcriteria for Hodgkin lymphoma and brain cancer aredescribed below and summarized in Table 1.

Hodgkin lymphomaSeventy-three peer-reviewed publications evaluatingHHV-6 in Hodgkin lymphoma were reviewed for thecriteria listed above. Eight studies with serologic datawere identified [69, 72–78].

Serologic studiesThe initial case-control serologic studies evaluating anti-bodies to HHV-6 in Hodgkin lymphoma were based onsamples of convenience and generally showed higherantibody titers in patients with HL compared to popula-tion controls. Ablashi et al. [72] used an immunofluores-cence assay beginning at a 1:20 dilution, finding only26% of healthy donors in the U.S. Canada and Europewere positive compared to 77.2% of patients with Hodgkinlymphoma [72]. Clark et al. examined sera from patientswith diverse hematologic malignancies which includedtwo groups of HD patients, using various dilutions of anti-body detected by immunofluorescence [73]. They ob-served elevated titers in HL but also in acute myeloidleukemia and low grade non-Hodgkin lymphoma [73].Other investigators, including Biberfeld et al. and Torelliet al., also found higher antibody titers in HL than in vari-ous control groups [74, 75].In contrast, Levine et al. obtained divergent results in

37 longitudinally-followed patients with HL in Denmarktested for HHV-6 by both immunofluorescence assay(IFA) and enzyme-linked immunosorbent assay (ELISA)[76]. The major findings in this study were: 1) Initial

pre-treatment samples showed lower HHV-6 antibodytiters in HL patients compared to controls. 2) ELISA andIFA results were discordant, with IFA rising and ELISAdeclining over time showing the importance of the assayused. 3) Antibodies showed major changes with treatmentin HL cases but were remarkably stable in healthy controlsstudied longitudinally. 4) Both EBV and HHV-6 IFA titersrose longitudinally in relapsed but not in patients whodid not relapse. No studies other than, Levine, et al.[76], have examined HHV-6 antibodies using prospectivecohort or retrospective-prospective cohorts, although suchapproaches provided important etiological data for EBV.

Virus detection studiesThe initial virus detection studies [75, 77, 78] looked forthe presence of HHV-6 in HL tissues, but only onedetermined the specific location of the virus in Reed-Sternberg (RS) cells, the tumor cell in HL [77]. HHV-6was not only found in HL but also in other lymphomasand in non-malignant reactive lymph nodes [79]. Onestudy by Torelli et al. [74] exemplified the pattern inseveral reports suggesting that while the percentage ofHL cases with HHV-6 detected in the lymph nodes wassmall, the possibility remained that a subset of casessuch as three young women with similar clinical andhistologic subtypes of HL reported as nodular sclerosis-lymphocyte depletion HL could be etiologically linked toHHV-6.As noted above, one major criterion for viral oncogen-

icity is finding evidence of the virus in every tumor cell,suggesting the virus was the precipitating oncogene car-ried along with every subsequent mitotic division. In herreview of viruses possibly related etiologically to HL, Jar-rett [69] summarized the HHV-6 virus detection studiesand noted that no report to date has documented thisfinding. A relevant paper of interest by Luppi et al. [80]looked for HHV-6 expression in a variety of benign andmalignant diseases, finding viral late antigen expressionin sinus histiocytosis with massive lymphadenopathy(SHML or Rosai-Dorfman disease) confirming our

Table 1 IARC criteria for oncogenicity for six oncoviruses with HHV-6 comparison

Criteria Characteristics EBV HTLV-1 HHV-8 HPV HBV HCV HHV-6

Major criteria Viral detection (diffuse) X X X X X

Tumor prevention (by antiviral vaccine) a+/− X

Minor criteria Serology X X X X X X

Viral detection (restricted) N/A N/A N/A N/A N/A X

Biological gradients X X X X X

Clustering X X X X b+/−

Medical interventions X X X X

Animal models X X XaVaccine reduces number of precancerous lesions [51]bSome evidence supports clustering of HHV-6 [70, 71]

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 5 of 12

original report in 1992 [81]. However, Luppi et al. foundHHV-6 was detected rarely in HL, and when it was iden-tified in Reed-Sternberg cells, the cells were degenerat-ing indicating a lytic process, not compatible with thecontinued perpetuation of the virus in constantly divid-ing malignant cells. Another intriguing paper by Siddonet al. [77] reported using immunohistochemistry, poly-merase chain reaction (PCR), and fluorescent in situhybridization (FISH) to localize the virus in Hodgkintissues. Lymph nodes from 21 HL patients, nodularsclerosis subtype, were selected and 18 (86%) wereHHV-6 positive with 10 of these containing HHV-6-positive RS cells. Four different monoclonal antibodies(to whole virus, HHV-6 p41 late antigen, p98 late anti-gen, and U94 latent antigen), as well as FISH, identifiedthe virus in the RS cells. The author’s note that the iden-tification of multiple copies of the virus detected byFISH is suggestive of viral replication in some RS cells,and photomicrographs document “numerous HHV-6positive Reed-Sternberg (RS) cells” but the percentage ofpositive RS cells was not recorded.

Biological gradientHHV-6 DNA has been identified in white blood cellsand plasma of pediatric HL patients [82], suggesting theopportunity for viral load studies. Thus far, however, wehave not identified any studies that evaluate biologic gra-dient (correlation of viral load with the presence andparticularly the stage of disease or DNA copy number inPCR detection studies of tumors).

ClusteringTime-space clustering, an intriguing observation suggest-ing an acute triggering event, is more likely to be revealingwith aggressive malignancies with a rapid doubling time,such as acute leukemia [83], Burkitt lymphoma [84, 85] orinflammatory breast cancer [86, 87]. HL is far less aggres-sive but other approaches to clustering and transmissibil-ity of an infectious agent have been reported and analyzed[70, 88]. The initial reports were particularly focused onhigh school students and the nodular sclerosis form of thislymphoma. The transmission of the proposed etiologicagent has been suggested as either being directly from pa-tient to patient or via a single clinically unaffected “carrier”and the median interval between the diagnosis of theinfective patient or clinically unaffected carrier and theinfected patient has been estimated as 3 years [70, 88].

Medical interventionsNo medical intervention studies were identified.

Animal studiesNo animal studies were identified.

Brain CancerEleven studies were found for brain cancer [67, 89–98].

Serologic studiesOne report by Cuomo et al. [95] showed no differencein the frequency of sera positive for HHV-6 in 73 pa-tients with brain tumors compared to 150 healthy sub-jects. Only patients with brain tumors, however, hadantibody titers > 1:320, suggesting that there could be asmall subset of patients with brain tumors triggered byHHV-6 or reacting anomalously to it.

Virus detection studiesHHV-6 has been shown to be neurotropic and patho-genicity for humans has been well documented [67].Since HHV-6 had demonstrated tropism for normal andneoplastic cells of neuronal and glial origin [97], it wasreasonable to investigate the presence of HHV-6 DNAsequences by PCR in brain tumor specimens. Early workon the association of HHV-6 with brain cancer wasstudied by older, less sophisticated PCR techniques.Luppi et al. used frozen and paraffin-embedded samplesfrom different regions of brains obtained at necropsyfrom 9 control immunocompetent adults and 37 casesof primary brain tumors of neuroglial origin [97]. Sixnormal brain tissues (66%) were positive for HHV-6whereas only 7 of 37 (19%) brain tumors were positive.Similarly, Chan et al. [96] used a nested HHV-6 PCRmethod from 110 formalin fixed, paraffin-embedded sur-gical biopsies of primary brain tumors and found only8.2% were positive (no non brain-tumor samples wereanalyzed).By contrast, Crawford and colleagues had a series of

studies noting HHV-6A and 6B in brain tumors of chil-dren and adults that focused on the possible role ofthese viruses in pediatric and adult gliomas [89, 90].These reports used an HHV-6 nested PCR methodology,immunohistochemistry, and in situ hybridization studies.HHV-6 was more often detected in the tumor samplescompared to age-matched non-tumor brains [89]. Ofinterest is the observation that the HHV-6A variant wasmore often detected. The authors appropriately notethat these studies are more associative than causal, anddo not suggest that HHV-6 is a direct cause of brain tu-mors. These early results were later supported by Chi,et al., [92] that investigated the frequency of herpesvi-ruses in 40 glioma tissues compared to 13 normal brainsby using both immunohistochemical (IHC) analysis andnested PCR. Of the 5 human herpesviruses tested (HSV-1, EBV, Cytomegalovirus (CMV), HHV-6 and HHV-7),only HHV-6 was detected significantly more frequentlyin gliomas (42.5%) than in normal brain (7.7%). More-over, these observations were further supported by posi-tive IHC staining for HHV-6 in a subset of the tumor

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 6 of 12

samples (32.5%) and never observed in normal brain tis-sue [92]. Of all the different types of glioma tissues, 50%of glioblastoma multiforme (GBM) specimens wereHHV-6 positive.Other studies on GBM have used a novel and highly

precise multiplexed droplet digital PCR to assay from112 brain tissue specimens that included 45 GBM and49 control brains [98]. Each tissue section was analyzedfor HHV-6A, HHV-6B, EBV and CMV. Neither CMVnor HHV-6A were detected in any of the brain samples,while HHV-6B and EBV had a higher frequency of de-tection in the GBM samples compared to controls [98].This cross-sectional survey of multiple human herpesvi-ruses in high and low grade astrocytomas, as with subse-quent studies, only showed that multiple herpes virusescan be detected in brain tumors.

Biological gradientNo studies were identified.

ClusteringAs with HL, the latent period between the initiation of abrain tumor and its clinical appearance is far longer thanwith those aggressive malignancies reported to having acloser pattern of time-space clustering to an infectiousdisease. Evidence for this long latent period is providedby studies of known oncogenic agents, such as radiation,and the onset of the brain tumor, as observed by the ex-perience in Israel where radiation for tinea capitis wasshown to cause subsequent brain cancer many yearslater [99, 100]. One observation, possibly relevant toHHV-6, was a transient increase in brain cancer followinga cluster of chronic fatigue syndrome (CFS) [101, 102] thatwas linked to HHV-6 [71].CFS clusters, originally designated epidemic neuro-

myasthenia [103], have been reported since 1959 [104]and can apparently be triggered by different etiologicagents [105–108]. We focused on this particular out-break in the Lake Tahoe region [109] affecting northernNevada and nearby California because the reportingphysicians believed they were now seeing a surge of can-cer patients in their practice. Preliminary studies wereconducted using the Nevada Cancer Registry and twomalignancies, non-Hodgkin lymphoma (NHL) and braincancer. These were all identified as occurring shortlyafter the CFS cluster in northern Nevada and were notseen in southern Nevada [106, 107]. The link betweenCFS and NHL was confirmed in a large study of CFSand cancer. This involved more than a million individ-uals greater than 65 years of age, combining two largedata bases from the Surveillance, Epidemiology and EndResults (SEER) program and Medicare. NHL was theonly cancer with a statistically significant link to CFS[110]. A subsequent review of an additional 10 years of

data from the Nevada Cancer Registry [111] found thatthe increase in brain cancer in northern Nevada remainedstatistically significant but was restricted to only thoseunder age 65 and was therefore were not observable in theNCI SEER-Medicare analysis.

Medical interventionsThere are few interventional clinical trials that have tar-geted HHV-6-associated cancers. However, there wasone controversial report that described 50 patients withGBM who received valganciclovir as an add-on to stand-ard therapy [112]. In the valganciclovir/GBM study,there was a higher rate of 2-year survival (62%) in GBMpatients treated with this anti-beta herpesvirus drug thancontemporary controls (18%) not receiving valganciclovir[112]. However, selection bias may have influenced theseresults [113], and therefore antiviral treatment in GBMshould be extended to a properly designed randomizedtrial.

Animal studiesA recent study on the role of HHV-6 in chronic progres-sive neurologic disease may help to inform our under-standing on how these viruses may act to trigger oraccelerate disease [114]. Recently, it has been reportedthat HHV-6 infection in marmosets, a non-human pri-mate, by intranasal route [115] resulted in a benign, qui-escent infection in which viral material was recoveredfrom peripheral compartments. Subsequent challenge ofthese animals with CNS antigens resulted in an auto-immune encephalomyelitis (EAE), which was moresevere in animals that had been infected with HHV-6than in uninfected animals in which EAE was induced[114]. Mechanistically, expansion of proinflammatoryCD8+ T cells correlated with post-EAE survival in thevirus inoculated animals suggesting that a peripheralantigen-driven expansion may have occurred [114].These studies suggested that viral infection with a ubi-quitous virus such as HHV-6 can ‘prime’ disease devel-opment in challenge models in which the challengealone rarely induces disease [116]. Moreover, peripheralCNS viral infections can result in an enhanced responseto subsequent autoimmune challenge and supports thehypothesis that viruses such as HHV-6 may act as trig-gers for inflammatory mediated neurologic disease oreven CNS tumors.

DiscussionThe criteria for causation have changed over the years.We have evaluated the available data on HHV-6 as ahuman carcinogen using the criteria for accepted viralcarcinogens such as EBV, HPV and HTLV-I by currentstandards and acceptance by IARC as a Group 1 carcin-ogens. For all of the viral carcinogens, EBV appears to

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 7 of 12

be the best model for HHV-6 as it is ubiquitous world-wide and both are herpesviruses.For our major criteria, only identification of the virus

in all of the tumor cells is relevant since HHV-6 infec-tion is not considered preventable. Attempts to apply anEBV vaccine to prevent specific tumors in high riskareas, such as NPC in southern China, have been consid-ered [117] but not yet applied. Thus far, HHV-6 has notbeen demonstrated in every tumor cell in either HL orbrain cancer. For Hodgkin lymphoma, EBV has been ac-cepted as the cause of up to 30% of cases by identificationof EBERs in every Reed-Sternberg cell [45, 46, 49, 69].HHV-6 is also sometimes found in Reed-Sternberg cellsbut only in occasional cells (restricted). HHV-6 has beenreported as present in all tumor cells (diffuse) in oralcancer, thus meeting this major criterion [118], but thissingle report remains unconfirmed.In regard to minor criteria, there are possible oppor-

tunities to examine whether HHV-6 meets any of thecriteria used to show EBV oncogenicity. Linking ahuman virus to cancer usually begins with serology. Asdescribed in an extensive review [119], elevated HHV-6titers have been observed in patients with numeroushuman tumors. For EBV, however, the titers have uni-versally been elevated in the sera of individuals destinedto get the tumors years to decades in advance of diagno-sis. In addition to Burkitt’s lymphoma [55] and gastriccancer [57], nasopharyngeal carcinoma is now an im-portant malignancy where EBV serology has proven ef-fective in screening of very high risk populations, suchas high risk families [120]. Further studies will improvethe specificity of assays that had earlier been used forlarger populations [76].As we have noted, all of the reports showing high

HHV-6 titers have been done in treated patients. To ourknowledge, the only study using sera from untreated HLpatients showed no elevation compared to controls (andindeed were somewhat lower than controls); in thatstudy, titers increased only after treatment [76]. Similarprospective longitudinal studies should be developedwith HHV-6, so that findings parallel to EBV findingscan be examined. It would be of value to use some ofthe large serum banks such as the prostate, lung, colo-rectal, and ovarian cancer screening trial (PLCO) wherelarge prospective studies collect blood and other bio-logical samples on clinically healthy individuals yearsbefore tumors were diagnosed to determine risk factorsfor subsequent disease, including cancer.Most of the other minor criteria such as viral load pre-

disease, medical intervention and animal models have notdemonstrated HHV-6 to be related to either Hodgkinlymphoma or brain cancer with the exception of thevalgancyclovir study in glial tumors (not yet confirmed).For both Hodgkin lymphoma and brain cancer, there have

been intriguing epidemiologic leads under the category ofclustering. As noted above, time-space clusters of cancerare only likely to be worthy of consideration with rareaggressive tumors which have a short latent period, suchas acute leukemia [83], Burkitt’s lymphoma [84, 85], andinflammatory breast cancer [86, 87]. Thus far, investiga-tion of these clusters has rarely, if ever, shown to have aspecific cause. An increase in a rare cancer not necessarilyin a cluster can be a fruitful finding. For example, investi-gations of vaginal cancer in young women showed an etio-logic role for diethylstilbestrol [121]. Mapping of cancermortality has been useful, successfully showing asbestosexposure, especially associated with smoking, as the causeof mesothelioma in shipyard workers [122]. Cancermortality mapping also led to studies showing that to-bacco chewing caused oral cancer in women in thesoutheast United States [123]. For specific viruses, themodel for cancer clustering linked to a virus is HTLV-I,which could be linked to adult T-cell leukemia/lymph-oma. It is not readily transmissible and therefore has alimited distribution.The studies of interest in Hodgkin lymphoma and

brain cancer are provocative because they indicate groupsof patients where studies should be focused in determin-ing a possible viral etiology. For Hodgkin lymphoma, theprominence of clustering of cases reported by Viannaet al. [124, 125] focus on a group that is most probablynot EBV-associated: high school students with nodularsclerosing HL. EBV is predominantly associated withHodgkin lymphoma in childhood and in the elderly andwith the histological characterization as mixed cellularityhistological subtype. For brain cancer, the increased inci-dence of brain tumors in northern Nevada [106, 107, 110]following an outbreak of chronic fatigue syndrome previ-ously linked to HHV-6 [71] may represent “clustering”.Attention should be given to identifying specific groupsfor intensive study rather than all patients with the tumor.This is particularly applicable to oral cancer, the onetumor where it has been suggested that the virus ispresent in every tumor cell but not normal cells [118].This study used tumors from an area in India with ele-vated HHV-6 titers in the healthy population. It could besuggested that attention should be given to patients inareas with elevated HHV-6 titers [63] as possibly beingmore fruitful than low titer populations.It should be noted that considerable attention is being

given to properties of HHV-6 consistent with oncogenicpotential such as oncomodulation and chromosomal in-tegration [67, 119]. These properties are also present innon-oncogenic viruses, however, and therefore at thistime we believe a focus on mechanisms of HHV-6 onco-genicity are premature.It is important to consider the viral differences be-

tween EBV and HHV-6, one being a gamma herpesvirus

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 8 of 12

and the other a beta herpesvirus, respectively. Each virushas biological and genetic differences, which may trans-late into different oncologic mechanisms that affect theacceleration or development of cancer. An in-depth de-scription of these differences is beyond the scope of thisreview.

ConclusionThere is insufficient evidence to indicate HHV-6 is anetiologic agent with respect to Hodgkin lymphoma andbrain cancers. Based on minor criteria, HHV-6 may playsome role in the development of these two malignancies.Studies report evidence for both HHV-6A and HHV-6B be-ing involved in HL [77–79] and brain cancer [89, 90, 92].Therefore, we suggest classifying both HHV-6A and HHV-6B as IARC Group 2B, possibly oncogenic for humans.We suggest that methods demonstrating EBV oncogen-

icity be applied to HHV-6. In addition, there is at leastone other cancer, oral cancer, where one of our major cri-teria has reportedly been fulfilled. Therefore HHV-6 canstill be considered as a possible human oncogenic virus.Although the evidence for HHV-6 as an etiologic

agent for Hodgkin lymphoma and brain cancer is cur-rently weak, there are two important opportunities tostrengthen the case for it being designated as a Class Icarcinogen by IARC. First, it could be helpful if some ofthe studies noted as supporting the role of EBV in onco-genesis, such as the development of prospective studiesin high-risk populations, are implemented for HHV-6.Second, additional studies of oral cancer in Indian popu-lations with elevated HHV-6 antibodies could confirmthe importance of the study indicating the presence ofHHV-6 in all of the tumor cells [118].

AbbreviationsATLL: Adult T-cell leukemia/lymphoma; BL: Burkitt lymphoma; EBV: Epstein-Barr virus; HAM: HTLV-I associated myelopathy; HBV: Hepatitis B virus;HBV: Hepatitis C virus; HHV-6: Human herpesvirus-6; HHV-7: Humanherpesvirus 7; HHV-8: Human herpesvirus 8; HIV: Human immunodeficiencyvirus; HL: Hodgkin lymphoma; HPV: Human papilloma virus; HSV-2: Herpessimplex virus 2; HTLV- I: Human T-cell lymphotropic Virus Type-I;IARC: International Agency for Research on Cancer; KSHV: Kaposi’s sarcomaherpesvirus; MCV: Merkel cell tumor virus; NPC: Nasopharyngeal carcinoma;PCR: Polymerase chain reaction; RS: Reed-sternberg; SHML: Sinus histiocytosiswith massive lymphadenopathy

AcknowledgementsThe authors thank Drs. Robert Biggar and Ruth Jarrett for helpful comments,and Emily Lum for technical assistance. The authors also thank Drs. AllanHildesheim and Charles Rabkin of the National Cancer Institute, Infectionsand Immunoepidemiology Branch, for suggesting the IARC criteria foroncogenesis and providing the background documents utilized for theviruses designated Class 1 carcinogens.

Authors’ contributionsAll authors reviewed the literature and participated in the writing andediting the manuscript. All authors read and approved the final manuscript.

FundingInitial support provided by the HHV-6 Foundation, which did not review orcontribute to this manuscript. SJ is supported by the intramural research pro-gram of the National Institutes of Health, NINDS.

Availability of data and materialsNot applicable

Ethics approval and consent to participateNot applicable

Consent for publicationNot applicable

Competing interestsThe authors declare that they have no competing interests. Partial supportfor this manuscript was provided by the HHV-6 Foundation which did notparticipate in the preparation and did not review the manuscript prior tosubmission.

Author details1School of Community and Population Health, University of New England,716 Stevens Ave, Portland, ME 04103, USA. 2National Institutes of Health,National Institute of Neurological Disorders and Stroke, Viral ImmunologySection, 9000 Rockville Pike, Bethesda, MD 20892, USA. 3College of PublicHealth, University of Nebraska, 984355 Medical Center, Omaha, NE 68198,USA.

Received: 7 August 2019 Accepted: 26 September 2019

References1. Braun DK, Dominguez G, Pellett PE. Human herpesvirus 6. Clin Microbiol

Rev. 1997;10(3):521–67.2. Readhead B, Haure-Mirande J-V, Funk CC, Richards MA, Shannon P,

Haroutunian V, et al. Multiscale Analysis of Independent Alzheimer’s CohortsFinds Disruption of Molecular, Genetic, and Clinical Networks by HumanHerpesvirus. Neuron. 2018;99(1):64–82.e7.

3. Campadelli-Fiume G, Mirandola P, Menotti L. Human herpesvirus 6: Anemerging pathogen. Emerg Infect Dis. 1999;5(3):353–66.

4. Challoner PB, Smith KT, Parker JD, MacLeod DL, Coulter SN, Rose TM.Plaque-associated expression of human herpesvirus 6 in multiple sclerosis.Proc Natl Acad Sci U S A. 1995;92:7440–4.

5. Salahuddin SZ, Ablashi DV, Markham PD, Josephs SF, Sturzenegger S, KaplanM, et al. Isolation of a new virus, HBLV, in patients with lymphoproliferativedisorders. Science. 1986;234(4776):596–601.

6. Lopez C, Pellett P, Stewart J, Goldsmith C, Sanderlin K, Black J, et al.Characteristics of human herpesvirus-6. J Infect Dis. 1988;157(6):1271–3.

7. Lusso P, Markham PD, Tschachler E, di Marzo VF, Salahuddin SZ, Ablashi DV,et al. In vitro cellular tropism of human B-lymphotropic virus (humanherpesvirus-6). J Exp Med. 1988;167(5):1659–70.

8. Takahashi K, Sonoda S, Higashi K, Kondo T, Takahashi H, Takahashi M, et al.Predominant CD4 T-lymphocyte tropism of human herpesvirus 6-relatedvirus. J Virol. 1989;63(7):3161–3.

9. Ablashi D, Agut H, Alvarez-Lafuente R, Clark DA, Dewhurst S, Diluca D, et al.Classification of HHV-6A and HHV-6B as distinct viruses. Arch Virol. 2014;159(5):863–70.

10. Arbuckle JH, Medveczky MM, Luka J, Hadley SH, Luegmayr A, DharamAblashi D, et al. The latent human herpesvirus-6A genome specificallyintegrates in telomeres of human chromosomes in vivo and in vitro. ProcNatl Acad Sci U S A. 2010;107(12):5563–8.

11. Suga S, Yazaki T, Kajita Y, Ozaki T, Asano Y. Detection of human herpesvirus6 DNAs in samples from several body sites of patients with exanthemsubitum and their mothers by polymerase chain reaction assay. J Med Virol.1995;46(1):52–5.

12. Suga S, Yoshikawa T, Kajita Y, Ozaki T, Asano Y. Prospective study ofpersistence and excretion of human herpesvirus-6 in patients withexanthem subitum and their parents. Pediatrics. 1998;102(4 Pt 1):900–4.

13. Baillargeon J, Piper J, Leach CT. Epidemiology of human herpesvirus 6 (HHV-6) infection in pregnant and nonpregnant women. J Clin Virol. 2000;16(3):149–57.

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 9 of 12

14. Blattner WA, Kalyanaraman VS, Robert-Guroff M, Lister TA, Galton DA, SarinPS, et al. The human type-C retrovirus, HTLV, in blacks from the Caribbeanregion, and relationship to adult T-cell leukemia/lymphoma. Int J Cancer.1982;30(3):257–64.

15. Hinuma Y. Antibodies to adult t-cell leukemia-virus-associated antigen (atla)in sera from patients with atl and controls in japan: A nation-wide sero-epidemiologic study. Int J Cancer. 2006;29:631–5.

16. Proietti F, Carneiro-Proietti A, Catalan-Soares B, Murphy E. Globalepidemiology of HTLV-1 infection and associated diseases. Oncogene.2005;24:6058–68.

17. Henle J. “On Miasmata and Contagie” (trans. With an introduction byGeorge Rosen). Baltimore: Johns Hopkins Press; 1938.

18. Koch R. The etiology of tuberculosis assistants from the Imperial healthoffice 1884;2:1–88.

19. Koch, R. About bacteriological research. Neg. X. Internat. Med. Congr. Berlin.1891 Bd. I: 35–47.

20. Hill AB. The environment and disease: association or causation? Proc R SocMed. 1965;58:295–300.

21. Rivers TM. Viruses and Koch’s postulates. J Bacteriol. 1937;33(1):1–12.22. Fredericks DN, Relman DA. Sequence-based identification of microbial

pathogens: a reconsideration of Koch's postulates. Clin Microbiol Rev.1996;9(1):18–33.

23. Moore PS, Chang Y. The conundrum of causality in tumor virology: thecases of KSHV and MCV. Semin Cancer Biol. 2014;26:4–12.

24. IARC. Schistosomes, liver flukes and Helicobacter pylori. IARC WorkingGroup on the Evaluation of Carcinogenic Risks to Humans Lyon. France:IARC monographs on the evaluation of carcinogenic risks to humans;1994. p. 1–241.

25. Graham DY. History of helicobacter pylori, duodenal ulcer, gastric ulcer andgastric cancer. World J Gastroenterol. 2014;20(18):5191–204.

26. Keck JW, Miernyk KM, Bulkow LR, Kelly JJ, McMahon BJ, Sacco F, et al.Helicobacter pylori infection and markers of gastric cancer risk in Alaskanative persons: a retrospective case-control study. Can J GastroenterolHepatol. 2014;28(6):305–10.

27. Martinez-Lopez JL, Torres J, Camorlinga-Ponce M, Mantilla A, Leal YA,Fuentes-Panana EM. Evidence of Epstein-Barr virus association with gastriccancer and non-atrophic gastritis. Viruses. 2014;6(1):301–18.

28. Zur Hausen A, van Rees BP, van Beek J, Craanen ME, Bloemena E, Offerhaus GJ,et al. Epstein-Barr virus in gastric carcinomas and gastric stump carcinomas: alate event in gastric carcinogenesis. J Clin Pathol. 2004;57(5):487–91.

29. Truong CD, Feng W, Li W, Khoury T, Li Q, Alrawi S, et al.Characteristics of Epstein-Barr virus-associated gastric cancer: a studyof 235 cases at a comprehensive cancer center in U.S.A. J Exp ClinCancer Res. 2009;28:14.

30. Jarrett RF, Gallagher A, Jones DB, Alexander FE, Krajewski AS, Kelsey A, et al.Detection of Epstein-Barr virus genomes in Hodgkin's disease: relation toage. J Clin Pathol. 1991;44(10):844–8.

31. Mueller BU, Butler KM, Higham MC, Husson RN, Montrella KA, Pizzo PA, et al.Smooth muscle tumors in children with human immunodeficiency virusinfection. Pediatrics. 1992;90(3):460–3.

32. Lee ES, Locker J, Nalesnik M, Reyes J, Jaffe R, Alashari M, et al. Theassociation of Epstein-Barr virus with smooth-muscle tumors occurring afterorgan transplantation. N Engl J Med. 1995;332(1):19–25.

33. IARC. Biological agents. A review of human carcinogens. IARC monographs onthe evaluation of carcinogenic risks to humans / World Health Organization,International Agency for Research on Cancer. 2012;100(B):1–441.

34. IARC. Kaposi's sarcoma herpesvirus/human herpesvirus 8. IARC monographson the evaluation of carcinogenic risks to humans / World HealthOrganization, International Agency for Research on Cancer. 1997;70:375–492.

35. IARC. Human T-cell lymphotropic viruses. IARC monographs on theevaluation of carcinogenic risks to humans / World Health Organization,International Agency for Research on Cancer 1996;67:261–390.

36. IARC. Hepatitis B virus. IARC monographs on the evaluation of carcinogenicrisks to humans / World Health Organization, International Agency forResearch on Cancer. 1994;59:45–164.

37. IARC. Hepatitis C virus. IARC monographs on the evaluation of carcinogenicrisks to humans / World Health Organization, International Agency forResearch on Cancer. 1994;59:165–221.

38. IARC. Human papillomaviruses. IARC monographs on the evaluation ofcarcinogenic risks to humans / World Health Organization, InternationalAgency for Research on Cancer. 1995;64:1–378.

39. IARC. Human papillomaviruses. IARC monographs on the evaluation ofcarcinogenic risks to humans / World Health Organization, InternationalAgency for Research on Cancer. 2007;90:1–636.

40. IARC. Merkel cell polyomavirus. IARC monographs on the evaluation ofcarcinogenic risks to humans / World Health Organization. InternationalAgency for Research on Cancer. 2014;104:309–50.

41. Tsai ST, Jin YT, Mann RB, Ambinder RF. Epstein-Barr virus detection innasopharyngeal tissues of patients with suspected nasopharyngealcarcinoma. Cancer. 1998;82(8):1449–53.

42. Shibata D, Hawes D, Stemmermann GN, Weiss LM. Epstein-Barr virus-associated gastric adenocarcinoma among Japanese Americans in Hawaii.Cancer Epidemiol Biomarkers Prev. 1993;2(3):213–7.

43. Shibata D, Tokunaga M, Uemura Y, Sato E, Tanaka S, Weiss LM. Associationof Epstein-Barr virus with undifferentiated gastric carcinomas with intenselymphoid infiltration. Lymphoepithelioma-like carcinoma. Am J Pathol. 1991;139(3):469–74.

44. Shibata D, Weiss LM. Epstein-Barr virus-associated gastric adenocarcinoma.Am J Pathol. 1992;140(4):769–74.

45. Weiss LM, Strickler JG, Warnke RA, Purtilo DT, Sklar J. Epstein Barr viral DNAin tissues of Hodgkin's disease. Am J Pathol. 1987;129(1):86–91.

46. Ambinder RF, Browning PJ, Lorenzana I, Leventhal BG, Cosenza H, Mann RB,et al. Epstein-Barr virus and childhood Hodgkin's disease in Honduras andthe United States. Blood. 1993;81(2):462–7.

47. Glaser SL, Lin RJ, Stewart SL, Ambinder RF, Jarrett RF, Brousset P, et al.Epstein-Barr virus-associated Hodgkin's disease: epidemiologic characteristicsin international data. Int J Cancer. 1997;70(4):375–82.

48. Sleckman BG, Mauch PM, Ambinder RF, Mann R, Pinkus GS, Kadin ME, et al.Epstein-Barr virus in Hodgkin's disease: correlation of risk factors and diseasecharacteristics with molecular evidence of viral infection. Cancer EpidemiolBiomarkers Prev. 1998;7(12):1117–21.

49. Evans AS, Mueller NE. Viruses and cancer. Causal Assoc Ann Epidemiol.1990;1(1):71–92.

50. Chang ET, Sue E, Zola J, So SK. 3 for life: a model pilot program to preventhepatitis B virus infection and liver cancer in Asian and Pacific islanderAmericans. Am J Health Promot. 2009;23(3):176–81.

51. Garland SM. The Australian experience with the human papillomavirusvaccine. Clin Ther. 2014;36(1):17–23.

52. Stone KM, Zaidi A, Rosero-Bixby L, Oberle MW, Reynolds G, Larsen S, et al.Sexual behavior, sexually transmitted diseases, and risk of cervical cancer.Epidemiol. 1995;6(4):409–14.

53. Henle G, Henle W, Clifford P, Diehl V, Kafuko GW, Kirya BG, et al. Antibodiesto Epstein-Barr virus in Burkitt's lymphoma and control groups. J NatlCancer Inst. 1969;43(5):1147–57.

54. Klein G, Geering G, Old LJ, Henle G, Henle W, Clifford P. Comparison of theanti-EBV titer and the EBV-associated membrane reactive and precipitatingantibody levels in the sera of Burkitt lymphoma and nasopharyngealcarcinoma patients and controls. Int J Cancer. 1970;5(2):185–94.

55. Geser A, de The G, Lenoir G, Day NE, Williams EH. Final case reporting fromthe Ugandan prospective study of the relationship between EBV andBurkitt's lymphoma. Int J Cancer J. 1982;29(4):397–400.

56. Beasley RP, Hwang LY, Lin CC, Chien CS. Hepatocellular carcinoma andhepatitis B virus. A prospective study of 22 707 men in Taiwan. Lancet.1981;2(8256):1129–33.

57. Levine PH, Stemmermann GN, Lennette ET, Hildesheim A, Shibata D,Nomura A. Elevated antibody titers to Epstein-Barr virus prior to thediagnosis of Epstein-Barr-virus-associated gastric adenocarcinoma. Int JCancer. 1995;60(5):642–4.

58. Nkrumah F, Henle W, Henle G, Herberman R, Perkins V, Depue R. Burkitt'slymphoma: its clinical course in relation to immunologic reactivities to Epstein-Barr virus and tumor-related antigens. J Natl Cancer Inst. 1976;57(5):1051–6.

59. Henle W, Ho JH, Henle G, Chau JC, Kwan HC. Nasopharyngeal carcinoma:significance of changes in Epstein-Barr virus-related antibody patternsfollowing therapy. Int J Cancer J. 1977;20(5):663–72.

60. Hisada M, Maloney EM, Sawada T, Miley WJ, Palmer P, Hanchard B, et al.Virus markers associated with vertical transmission of human Tlymphotropic virus type 1 in Jamaica. Clin Infect Dis. 2002;34(12):1551–7.

61. Uchiyama T, Yodoi J, Sagawa K, Takatsuki K, Uchino H. Adult T-cell leukemia:clinical and hematologic features of 16 cases. Blood. 1977;50(3):481–92.

62. Whitby D, Luppi M, Barozzi P, Boshoff C, Weiss RA, Torelli G. Humanherpesvirus 8 seroprevalence in blood donors and lymphoma patients fromdifferent regions of Italy. J Natl Cancer Inst. 1998;90(5):395–7.

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 10 of 12

63. Levine PH, Neequaye J, Yadav M, Connelly R. Geographic/ethnicdifferences in human herpesvirus 6 antibody patterns. MicrobiolImmunol. 1992;36:169–72.

64. Dewan M, Terashima K, Taruishi M. Rapid tumor formation of humanT-cell leukemia virus type 1-infected cell lines in novel NOD-SCID/γcnull mice: suppression by an inhibitor against NF-κB. J Virol. 2003;77(9):5286–94.

65. Epstein MA. Historical background. Philos Trans R Soc Lond Ser B Biol Sci.2001;356(1408):413–20.

66. Hardy WD, Zuckerman E. Ten-year study comparing enzyme-linkedimmunosorbent assay with the immunofluorescent antibody test fordetection of feline leukemia virus infection in cats. J Am Vet Med Assoc.1991;199(10):1363–73.

67. Kofman A, Marcinkiewicz L, Dupart E. The roles of viruses in brain tumor1000 initiation and oncomodulation. J Neuro-Oncol. 2011;105(3):451–66.

68. Quadrelli C, Barozzi P, Riva G. β-HHVs and HHV-8 in LymphoproliferativeDisorders. Mediterr J Hematol Infect Dis. 2011;3(1):e2011043.

69. Jarrett R. In: Engert A, Horning SJ, editors. Hodgkins Lymphoma: AComprehensive Update on Diagnostics and Clinics. 1st ed. BerlinHeidelberg: Springer-Verlag; 2011.

70. Vianna NJ, Polan AK. Epidemiologic evidence for transmission of Hodgkn’sdisease. N Engl J Med. 1973;289(10):499–502.

71. Buchwald D, Cheney P, Peterson D, Henry B, Wormsley S, Geiger A, et al. Achronic illness characterized by fatigue, neurologic and immunologicdisorders, and active human herpesvirus type 6 infection. Ann Intern Med.1992;116(2):103–13.

72. Ablashi DV, Josephs SF, Buchbinder A, Hellman K, Nakamura S, Llana T, et al.Human B-lymphotropic virus (human herpesvirus-6). J Virol Methods.1988;21(1–4):29–48.

73. Clark DA, Alexander FE, McKinney PA, Roberts BE, O'Brien C, Jarrett RF, et al.The seroepidemiology of human herpesvirus-6 (HHV-6) from a case-controlstudy of leukaemia and lymphoma. Int J Cancer. 1990;45(5):829–33.

74. Torelli G, Marasca R, Luppi M, Selleri L, Ferrari S, Narni F, et al. Human herpesvirus-6 in human lymphomas: identification of specific sequences in Hodgkin’slymphomas by polymerase chain reaction. Blood. 1991;77(10):2251–8.

75. Biberfeld P, Petren AL, Eklund A, Lindemalm C, Barkhem T, Ekman M, et al.Human herpesvirus-6 (HHV-6, HBLV) in sarcoidosis and lymphoproliferativedisorders. J Virol Methods. 1988;21(1–4):49–59.

76. Levine PH, Ebbesen P, Ablashi DV, Saxinger WC, Nordentoft A, Connelly RR.Antibodies to human herpes virus −6 and clinical course in patients withHodgkin’s disease. Int J Cancer. 1992;51:53–7.

77. Siddon A, Lozovatsky L, Mohamed A, Hudnall SD. Human herpesvirus 6positive reed-Sternberg cells in nodular sclerosis Hodgkin lymphoma. Br JHaematol. 2012;158(5):635–43.

78. Lacroix A. HHV-6 and EBV DNA quantitation in lymph nodes of 86 patientswith Hodgkin’s lymphoma. J Med Virol. 2007;79(9):1349–56.

79. Valente G, Secchiero P, Lusso P, Abete MC, Jemma C, Reato G, et al. Humanherpesvirus 6 and Epstein-Barr virus in Hodgkin's disease: a controlled studyby polymerase chain reaction and in situ hybridization. Am J Pathol. 1996;149(5):1501–10.

80. Luppi M, Barozzi P, Garber R, Maiorana A, Bonacorsi G, Artusi T, et al.Expression of human herpesvirus-6 antigens in benign and malignantlymphoproliferative diseases. Am J Pathol. 1998;153(3):815–23.

81. Levine PH, Jahan N, Murari P, Manak M, Jaffe ES. Detection of humanherpesvirus 6 in tissues involved by sinus histiocytosis with massivelymphadenopathy (Rosai Dorfman disease). J Infect Dis. 1992;166:291–5.

82. Loutfy SA, Fawzy M, El-Wakil M, Moneer MM. Presence of human herpesvirus 6 (HHV6) in pediatric lymphomas: impact on clinical course andassociation with cytomegalovirus infection. Virol J. 2010;7:287.

83. Heath CW, Hasterlik R. Leukemia amoung children in a suburbancommunity. Am J Med. 1963;34:796–812.

84. Morrow RH, Pike MC, Smith PG, Ziegler JL, Kisuule A. Burkitt’s lymphoma: atime-space cluster of cases in Bwanba County of Uganda. Br Med J. 1971;2(5760):491–2.

85. Levine PH, Sandler SG, Komp DM, O'Conor GT, O'Connor DM. Simultaneousoccurrence of "American Burkitt's lymphoma" in neighbors. NEJM. 1973;288(11):562-63.

86. Levine PH, Veneroso C. The epidemiology of inflammatory breast cancer.Semin Oncol. 2008;35(1):11–6.

87. Duke T, Jahed N, Veneroso C, Da Roza R, Johnson O, Hoffman D, et al. Acluster of inflammatory breast Cancer (IBC) in an office setting: additional

evidence of the importance of environmental factors in IBC etiology. OncolRep. 2010;24:1277–84.

88. Smith PG, Pike MC. Case clustering in Hodgkin’s disease: a brief review ofthe present position and report of current work in Oxford. Cancer Res. 1974;34:1156–60.

89. Crawford JR, Santi MR, Thorarinsdottir HK, Cornelison R, Rushing EJ, ZhangH, et al. Detection of human Herpesvirus-6 variants in pediatric braintumors: Association of Viral Antigen in low grade gliomas. J Clin Virol. 2009;46(1):37–42.

90. Crawford J, Santi M, Cornelison R, Sallinen S, Haapasalo H, MacDonald T.Detection of human herpesvirus-6 in adult central nervous system tumors:predominance of early and late viral antigens in glial tumors. J Neuro-Oncol. 2009;95(1):49–60.

91. Yoshikawa TAY, Akimoto S, Ozaki T, Iwasaki T, Kurata T, Goshima F,Nishiyama Y. Latent infection of human herpesvirus 6 in astrocytoma cellline and alteration of cytokine synthesis. J Med Virol. 2002;66(4):497–505.

92. Chi JGB, Zhang C, Peng G, Zhou F, Chen Y, Zhang G, Guo Y, Guo D, Qin J,Wang J, Li L, Wang F, Liu G, Xie F, Feng D, Zhou H, Huang X, Lu S, Liu Y, HuW, Yao K. Human herpesvirus 6 latent infection in patients with glioma. JInfect Dis. 2012;206(9):1394–8.

93. Amirian ESSM. Chromosomally-integrated human herpesvirus 6 in familialglioma etiology. Med Hypotheses. 2012;79(2):193–6.

94. Zheng XLS, Zang Z, Hu J, An J, Pei X, Zhu F, Zhang W, Yang H. Evidence forpossible role of toll-like receptor 3 mediating virus-induced progression ofpituitary adenomas. Mol Cell Endocrinol. 2016;5(426):22–32.

95. Cuomo L, Trivedi P, Cardillo M, Gagliardi F, Vecchione A, Caruso R, et al.Human herpes- virus 6 infection in neoplastic and normal brain tissue. JMed Virol. 2001;63:45–51.

96. Chan PKNH, Cheng AF. Detection of human herpesviruses 6 and 7 genomicsequences in brain tumours. J Clin Pathol. 1999;52(8):620–3.

97. Luppi MBP, Maiorana A, Marasca R, Trovato R, Fano R, Ceccherini-Nelli L,Torelli G. Human herpesvirus-6: a survey of presence and distribution ofgenomic sequences in normal brain and neuroglial tumors. J Med Virol.1995;47(1):105–11.

98. Lin C-TM, Emily C, Leibovitch MIA-S, Jacobson S. Human herpesvirusmultiplex ddPCR detection in brain tissue from low- and high-gradeastrocytoma cases and controls. Infect Agents Cancer. 2016;11:32.

99. Ron E, Modan B, Boice JJ, Alfandary E, Stovall M, Chetrit A, et al. Tumors ofthe brain and nervous system after radiotherapy in childhood. N Engl JMed. 1998;319(16):1033–9.

100. Sadetzki S, Chetrit A, Freedman L, Stovall M, Modan B, Novikov I. Long-termfollow-up for brain tumor development after childhood exposure toionizing radiation for tinea capitis. Radiat Res. 2005;163(4):424–32.

101. Holmes GP, Kaplan JE, Stewart JA, Hunt B, Pinsky JF, Shonberger LB. A clusterof patients with a mononucleosis-like syndrome. JAMA. 1987;257:2297–302.

102. Levine PH, Fears TR, Cummings P, Hoover RN. Cancer and a fatiguing illnessin northern Nevada--a causal hypothesis. Ann Epidemiol. 1998;8(4):245–9.

103. Henderson DA, Shelekov A. Epidemic neuromyasthenia—clinical syndrome?N Engl J Med. 1959;260:757–64.

104. Briggs NC, Levine PH. A comparative review of systemic and neurologicalsymptomatology in 12 outbreaks collectively described as chronic fatiguesyndrome, epidemic neuromyasthenia, and myalgic encephalomyelitis. ClinInfect Dis. 1994;18(1):S32–42.

105. Salit IE. Precipitating factors for the chronic fatigue syndrome. J PsychiatrRes. 1997;31(1):59–65.

106. Levine PH, Jacobson S, Pocinki AG, Cheney P, Peterson D, Connelly RR, et al.Clinical, epidemiologic, and virologic studies in four clusters of the chronicfatigue syndrome. Arch Intern Med. 1992;152:1611–6.

107. Levine PH, Peterson D, McNamee FL, O'Brien K, Gridley G, Hagerty M, et al.Does chronic fatigue syndrome predispose to non Hodgkin's lymphoma?Cancer Res. 1992;52:5516s–8s.

108. Levine PH, Snow PG, Ranum BA, Paul C, Holmes MJ. Epidemicneuromyasthenia and chronic fatigue syndrome in West Otago. NZ ArchIntern Med. 1997;157:750–4.

109. Barnes DM. Mystery disease at Lake Tahoe challenges virologists andclinicians. Science. 1986;234(4776):541–2.

110. Chang CM, Warren JW, Engels EA. ME/CFS associated with later life non-Hodgkin lymphoma: Medicare statistics. Cancer. 2012;118(23):5929–36.

111. Cheng-Te ML, Younes N, Levine PH. Incidence rates of brain cancerfollowing an outbreak of Chronic Fatigue Syndrome. J Epidemiol Res. 2018;5(1):1-6.

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 11 of 12

112. Soderberg-Naucler C, Rahbar A, Stragliotto G, Institutet K. Survival inpatients with glioblastoma receiving Valganciclovir. N Engl J Med. 2013;369(10):985–6.

113. Cobbs CS. Does valganciclovir have a role in glioblastoma therapy? Neuro-Oncol. 2014;16(3):331.

114. Leibovitch EC, Breanna C, Ha SK, Schindler MK, Lee NJ, Luciano NJ, et al.Herpesvirus trigger accelerates neuroinflammation in a nonhuman primatemodel of multiple sclerosis. Proc Natl Acad Sci U S A. 2018;115(44):11292–7.

115. Harberts E, Yao K, Wohler JE. Human herpesvirus-6 entry into the centralnervous system through the olfactory pathway. Proc Natl Acad Sci U S A.2011;108:13734–9.

116. Libbey JE, Fujinami RS. Potential triggers of MS. Results Probl CellDiffereration. 2010;51:21–42.

117. Pearson GR, Levine PH. Epstein Barr virus vaccine-the time to proceed isnow! Roizman B, Whitley R, Lopez C, editors. New York: Raven Press; 1993.

118. Yadav M, Chanrashekran DM, Ablashi D. Frequent Detection of HumanHerpesvirus 6 in Oral Carcinoma. J Natl Cancer Inst. 1994;86(23):1792–4.

119. Eliassen E, Lum E, Pritchett J, Ongradi J, Krueger G, Crawford JR, et al.Human herpesvirus 6 and malignancy: a review. Front Oncol. 2018;8:512.

120. Coghill A, Hsu WL, Pfeiffer RM, Juwana HH, Yu KJ, Lou PJ, et al. Epstein-Barrvirus serology as a potential screening marker for nasopharyngealcarcinoma among high-risk individuals from multiplex families in Taiwan.Cancer Epidemiol Biomarkers Prev. 2014;23(7):1213–9.

121. Herbst AL, Scully RE. Adenocarcinoma of the vagina in adolescence. Areport of 7 cases including 6 clear-cell carcinomas (so-calledmesonephromas). Cancer. 1970;25(4):745–57.

122. Blot WJ, Harrington JM, Toledo A, Hoover R, Heath CW, Fraumeni JF. Lungcancer after employment in shipyards during world war II. N Engl J Med.1978;299(12):620–4.

123. Winn DM, Blot WJ, Shy CM, Pickle LW, Toledo A, Fraumeni JF. Snuff dippingand oral cancer among women in the southern United States. N Engl JMed. 1981;304(13):745–9.

124. Vianna NJ, Greenwald P, Davies JNP. Nature of Hodgkin’s disease agent.Lancet. 1971;297(7702):733–36. https://doi.org/10.1186/s13027-019-0248-3.

125. Vianna NJ, Greenwald P, Davies JNP. Extended epidemic of Hodgkin’sdisease in high school students. Lancet. 1971;1209:10.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Wells et al. Infectious Agents and Cancer (2019) 14:31 Page 12 of 12